EESC6368 - Multimodal Signal Processing

EESC 6368 Multimodal Signal Processing (3 semester credit hours) Theory and applications in the field of multimodal signal processing. Robustness and performance of systems by considering cross-modal integration. Statistical algorithms and machine learning methods used for fusion/fission of multimodal content at feature, decision and model level. Common graphical models used in multimodal analysis including Dynamic Bayesian Network, Product Hidden Markov Model (HMM), Multistream HMM, Coupled HMM, Factorial HMM, Input Output HMM and segmental models. Recommended Corequisite: EESC 6349. Prerequisite: ENGR 3341 or equivalent. (3-0) T